Bacterial persisters tolerate antibiotic treatment, and underlie the propensity of biofilm infections to relapse. An improved understanding of persister physiology will lead to the development of more effective therapies against biofilm-utilizing pathogens such as Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus. Persister metabolism is of particular importance, as it influences entry into, maintenance of, and exit from this antibiotic-tolerant state. Unfortunately, persisters are a minor, transient subpopulation whose physiology is easily obscured by more abundant phenotypes (e.g., viable but non-culturable cells (VBNCs)). Recent evidence suggests that current persister isolation techniques provide samples with many more VBNCs than persisters. Without improved isolation techniques, the distinguishing feature between VBNCs and persisters, growth- resumption on standard media, must be used to characterize persister physiology. Here we propose to develop a method to elucidate the metabolic abilities of persisters from survival data, and thus circumvent the present isolation difficulties. Recent work has demonstrated that persisters can catabolize carbon sources, remain non-replicative, and yet become susceptible to aminoglycosides. Here we propose to harness this phenomenon to chart the metabolic capabilities of persisters. To accomplish this goal, we will develop a rapid AG potentiation assay, develop a computational approach to analyze the resulting survival data and direct further experimentation, and finally, demonstrate the utility of our approach by charting the metabolic abilities of three persister populations. To develop a rapid AG potentiation assay, we will use phenotype arrays (each well contains a different nutrient) to simultaneously measure AG potentiation and the necessary controls from hundreds of separate nutrients. To analyze the resulting survival data, we will use mixed integer linear optimization to generate an ensemble of non-redundant minimal metabolic pathways capable of explaining the data. These pathways will be clustered, and the reactions that most significantly discriminate between competing clusters will be experimentally perturbed to determine the metabolic capabilities of persisters. To demonstrate the utility of our approach, we will use our technique to study the metabolic abilities of three persister populations: exponential phase persisters tolerant to ofloxacin, exponential phase persisters tolerant to ampicillin, and stationary persisters tolerant to ofloxacin. Results from this proposal will fill fundamental knowledge gaps in persister metabolism, identify new avenues for therapeutic intervention through disruption of persister maintenance or enhancement of persister awakening, and impact the fields of network biology, microbiology, and infectious disease.